Load Balancing And Redundancy

Category: Routing and Switching
Type: Network Concepts
Generated on: 2025-07-10 09:03:38
For: Network Engineering, Administration & Technical Interviews

1. Quick Overview

What is it? Distributing network traffic across multiple servers or links to prevent any single point of failure and improve performance, availability, and scalability. Redundancy ensures service continuity in case of component failure.
Why is it important?
- Improved Performance: Distributes load, reducing the burden on individual servers.
- High Availability: Ensures service remains operational even if one server fails.
- Scalability: Allows easy addition of resources to handle increased traffic.
- Fault Tolerance: Minimizes downtime and service disruptions.

2. Key Concepts

Load Balancer: A dedicated device or software that distributes network traffic across multiple backend servers.
Server Pool (Backend Pool): A group of servers that provide the same service.
Health Check: Monitoring the status of backend servers to determine their availability.
Sticky Sessions (Session Persistence): Ensuring that a user’s requests are always directed to the same server. Crucial for applications that maintain state on the server.
High Availability (HA): A system designed to remain operational with minimal downtime, even in the face of failures.
Failover: The automatic switching to a redundant or standby system upon the failure of the primary system.
Active-Active: All servers in a cluster are actively processing requests.
Active-Passive: One server is active, and the other is on standby, ready to take over if the active server fails.
VIP (Virtual IP Address): The IP address that clients connect to. The load balancer maps the VIP to the actual IP addresses of the backend servers.
ECMP (Equal-Cost Multi-Path routing): A routing strategy where traffic is distributed across multiple paths with the same cost to the destination.
CARP (Common Address Redundancy Protocol): Allows multiple hosts on the same network to share an IP address. One host is the master, and the others are backups.
VRRP (Virtual Router Redundancy Protocol): Similar to CARP, but a standardized protocol.
Link Aggregation (LAG/EtherChannel): Combining multiple physical links into a single logical link to increase bandwidth and provide redundancy.
Server Affinity (Client Affinity): Directing traffic from the same client to the same server.

3. How It Works

Load Balancing (Example: HTTP)

Client sends request:
```
Client --> VIP (Load Balancer)
```
Load Balancer receives request:
- Applies load balancing algorithm (e.g., Round Robin, Least Connections).
- Performs health check on backend servers.
Load Balancer forwards request to a healthy server:
```
Load Balancer --> Server (e.g., Server 1)
```
Server processes request and sends response:
```
Server 1 --> Load Balancer
```
Load Balancer forwards response to client:
```
Load Balancer --> Client
```

ASCII Diagram:

    +--------+      +----------------+      +----------+
    | Client |----->| Load Balancer  |----->| Server 1 |
    +--------+      +----------------+      +----------+
                    |                |      | Server 2 |
                    | Health Checks  |----->| Server 3 |
                    +----------------+      +----------+

Redundancy (Example: Active-Passive with VRRP)

VRRP Configuration: Two routers (Router A and Router B) are configured with the same virtual IP address (VIP).
Router A is the Master: Router A has a higher priority and assumes the role of the Master. It owns the VIP.
Router B is the Backup: Router B listens for VRRP advertisements from Router A.
Traffic flows through Router A: All traffic destined for the VIP is routed to Router A.
Router A fails: Router B stops receiving VRRP advertisements from Router A.
Router B becomes the Master: After a timeout period, Router B assumes the role of Master and takes ownership of the VIP.
Traffic flows through Router B: All traffic destined for the VIP is now routed to Router B.

ASCII Diagram:

     +--------+       +--------+        +--------+
     | Client |------>|Router A|-------->|Network | (Master - Active)
     +--------+       |        |
                      | VRRP   |
     +--------+       |        |        +--------+
     | Client |------>|Router B|-------->|Network | (Backup - Passive)
     +--------+       +--------+        +--------+
                      (becomes Master on failure of A)

4. Protocol Details

VRRP (Virtual Router Redundancy Protocol - RFC 5798)
- Header:
  - Version
  - Virtual Router ID (VRID)
  - Priority
  - Advertisement Interval
  - Checksum
  - IP Address(es)
- Message Types:
  - Advertisement: Sent by the Master router to announce its presence and priority.
CARP (Common Address Redundancy Protocol): (OpenBSD)
- Similar to VRRP but simpler. Not standardized. Uses IP protocol 112.

5. Real-World Examples

E-commerce Website: Load balancing HTTP requests across multiple web servers to handle high traffic during peak shopping seasons. Using sticky sessions to maintain user shopping carts.
Database Clustering: Using replication and failover mechanisms to ensure database availability even if one database server fails.
Content Delivery Network (CDN): Distributing content across multiple servers geographically to reduce latency for users.
DNS Load Balancing: Using DNS to distribute traffic across multiple servers based on geographic location or other criteria.
Network Core Redundancy: Employing redundant core switches and routers using protocols like VRRP/HSRP and ECMP to ensure network backbone availability.
Cloud Services: Cloud providers heavily rely on load balancing and redundancy to provide scalable and highly available services.

6. Common Issues

Session Persistence Issues: If sticky sessions are not configured correctly, users may lose their session data.
- Solution: Verify sticky session configuration on the load balancer. Consider using a distributed session store (e.g., Redis, Memcached).
Health Check Failures: Backend servers may be incorrectly marked as unhealthy.
- Solution: Check the health check configuration on the load balancer. Ensure that the health check is appropriate for the application. Verify firewall rules are not blocking health checks.
Asymmetric Routing: Traffic may enter through one load balancer and exit through another, causing problems.
- Solution: Ensure that all load balancers are configured to handle both ingress and egress traffic. Consider using Direct Server Return (DSR).
Split-Brain Scenario (VRRP/CARP): Both routers may become the Master, leading to inconsistent routing.
- Solution: Ensure proper configuration of VRRP/CARP priorities and advertisement intervals. Implement a tie-breaker mechanism (e.g., using a management network).
Load Balancer Overload: The load balancer itself may become a bottleneck.
- Solution: Scale the load balancer by adding more resources or using multiple load balancers.

7. Configuration Examples

(a) HAProxy (Load Balancer - Linux)

frontend http-in
    bind *:80
    mode http
    default_backend webservers

backend webservers
    balance roundrobin
    server web1 192.168.1.101:80 check
    server web2 192.168.1.102:80 check
    server web3 192.168.1.103:80 check

(b) keepalived (VRRP - Linux)

vrrp_instance VI_1 {
    state MASTER
    interface eth0
    virtual_router_id 51
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass mypassword
    }
    virtual_ipaddress {
        192.168.1.100/24
    }
}

On the backup server, change:

state BACKUP
priority 90

(c) Cisco Router - HSRP (Hot Standby Routing Protocol - Cisco Proprietary, similar to VRRP)

interface GigabitEthernet0/0
 ip address 192.168.1.1 255.255.255.0
 standby version 2
 standby 1 ip 192.168.1.254  (Virtual IP)
 standby 1 priority 110        (Higher priority for Master)
 standby 1 preempt              (Preempt lower priority router)
 standby 1 track GigabitEthernet0/1 10  (Track interface for failover)

8. Interview Questions

What is load balancing, and why is it important? (See Quick Overview)
Explain the different load balancing algorithms you know.
- Round Robin: Distributes requests sequentially to each server.
- Least Connections: Sends requests to the server with the fewest active connections.
- IP Hash: Hashes the client’s IP address to determine which server to use. Ensures the same client always goes to the same server.
- Least Response Time: Sends requests to the server with the fastest response time.
- Weighted Round Robin: Distributes requests based on pre-defined weights assigned to each server.
What are sticky sessions (session persistence), and when are they necessary? (See Key Concepts). Give an example.
What is a health check, and why is it important? (See Key Concepts)
Explain the difference between active-active and active-passive redundancy. (See Key Concepts)
What is VRRP/HSRP, and how does it work? (See Key Concepts and How It Works)
How do you troubleshoot load balancing issues? (See Common Issues)
What are the security considerations when implementing load balancing?
- SSL termination: Load balancers can terminate SSL connections, reducing the load on backend servers. Ensure proper SSL certificate management.
- Protecting the load balancer itself: Implement security measures to protect the load balancer from attacks (e.g., DDoS).
- Rate limiting: Limit the number of requests from a single client to prevent abuse.
- Web application firewall (WAF): Protect against common web application attacks (e.g., SQL injection, XSS).

Detailed Answer Example (VRRP/HSRP):

VRRP (Virtual Router Redundancy Protocol) and HSRP (Hot Standby Routing Protocol - Cisco proprietary) are redundancy protocols that allow multiple routers to share a virtual IP address (VIP). One router is designated as the “Master” (or “Active” in HSRP) and is responsible for forwarding traffic destined for the VIP. The other routers are in a “Backup” (or “Standby” in HSRP) state and monitor the Master. If the Master fails, one of the Backup routers will take over as the new Master, ensuring continuous network connectivity. The Master router sends periodic “advertisement” messages to the Backup routers to indicate that it is still alive. VRRP is an open standard (RFC 5798), while HSRP is Cisco proprietary. VRRP is often preferred due to its open nature and wider support.

9. Related Concepts

Content Delivery Networks (CDNs): Use load balancing and redundancy to distribute content globally.
DNS (Domain Name System): Can be used for basic load balancing (DNS Round Robin).
Virtualization: Load balancing and redundancy are crucial in virtualized environments.
Cloud Computing: Cloud providers heavily rely on load balancing and redundancy to provide scalable and highly available services.
Network Security: Load balancing can be used to distribute traffic across multiple security devices (e.g., firewalls, intrusion detection systems).
Traffic Shaping/QoS (Quality of Service): Prioritizing certain types of traffic to improve performance.
Network Monitoring: Essential for identifying and resolving load balancing and redundancy issues.

This comprehensive cheatsheet provides a solid foundation for understanding and implementing load balancing and redundancy in computer networks. Remember to practice with real-world scenarios and lab environments to solidify your knowledge.